Worked on the crewAIInc/crewAI repository, focusing on improving the reliability of experiment workflows by addressing a critical bug in the Crew Template’s parameter handling. Applied Python development and CLI skills to correct the current_year parameter name and ensure its proper propagation to both training and testing flows. This targeted fix enhanced the robustness of template pipelines, preventing data window mismatches and improving reproducibility across experiments. The work involved careful debugging and code quality improvements within the template handling logic, resulting in more dependable data integrity for downstream processes. All changes were implemented in Python, emphasizing maintainable and accurate parameter management.